Abstract

Implementing machine learning in law would transform current legal orders, based on the rule of law. The result would be “legal singularity”: an order based on “precisely tailored laws, specifying the exact behaviour that is permitted in every situation”. According to its proponents, this would promote justice and legal certainty. Through a comparison with the Platonic proposal of the philosopher-king, this article defends that, even if the aforementioned values were to be promoted, the inherent opacity of machine learning systems would hamper the public scrutiny of the resulting normative order, affecting its legitimacy. Thus, current political communities must deliberate about the benefits and drawbacks of legal singularity and must reflect on the use of artificial intelligence in law.